Drought-Induced Decrease in Soil Respiration: A Transient Soil Carbon Sink?

Low rates of soil respiration during droughts have been related to above-average rates of net ecosystem exchange of carbon in forested ecosystems. Lower soil respiration during drought could cause a transitory soil C sink, due to a temporary decline in decomposition, or it could result from reduced root respiration. A throughfall exclusion experiment was established at the Harvard Forest in central Massachusetts to study the effects of experimental drought on soil respiration. Weekly measurements of soil respiration began in the spring of 2001, with 4 manual CO₂ flux chambers installed in each of 6 plots (5 x 5 m). In July 2001, sub-canopy roofs with translucent plastic panels were installed in 3 of the plots, while the other 3 control plots were left open. Temperature probes, TDR probes, and gas tubes were buried at 4 depths in each plot. Roofs were removed in autumn 2001 to allow leaf-fall and snowfall, reinstalled in spring 2002, removed again in autumn 2002, and then left off for the entire summer of 2003. The radiocarbon contents of CO₂ emissions and concentrations within the soil were measured periodically. Soil respiration was nearly identical in treatment and control plots (140 and 142 g C m^-2, respectively) during the 54-d pre-treatment period. During 85 days in 2001 with roofs in place, 168 mm of throughfall were diverted, and the cumulative soil respiration was 30% lower in exclusion versus control plots (237 and 338 g C m^-2, respectively). During 127 days of throughfall exclusion (344 mm diverted) in 2002, cumulative soil respiration was 40% lower in exclusion versus control plots (280 and 490 g C m^-2, respectively). Clearly, soil water content is an important factor influencing soil respiration. When natural throughfall was allowed in all plots from September 2002 through August 2003, the cumulative soil respiration was only 6% higher in exclusion versus control plots (702 and 658 g C m^-2, respectively), indicating only a modest increase in decomposition after previously dried soils were wetted. The Δ ¹⁴CO₂ in the soil atmosphere was negatively correlated with CO₂ concentrations, indicating a larger contribution of decomposition of old, radiocarbon-rich substrates in the mineral soil during drought. In dry mineral soil, root respiration produced less CO₂, while gradual decomposition of old, radiocarbon-rich substrates continued at low rates, resulting in low CO₂ concentrations with high Δ ¹⁴C values. In contrast, microbial decomposition of old substrates declined more rapidly than did decomposition of young substrates and root respiration in the O horizon, causing the Δ ¹⁴CO₂ of surfaces fluxes to decline with drought. A depth-dependent differential response of root and microbial respiration to drought is indicated. If decreased root respiration is responsible, it may reflect altered belowground C allocation, but larger treatment plots that include entire trees would be needed to address this question. In any case, a transient soil C sink explains only a small fraction of the drought-induced reduction in soil respiration in this experiment.